DysReg: Identify gene dysregulations

In SCBIT-YYLab/DysRegSig: Gene dysregulation analysis and construction of mechanistic signature in cancer

Description

Identify gene dysregulations by integrating three properties including differential regulation, differential expression of target, and the consistency between differential regulation and differential expression. DysReg could consider the combinational effect of multiple regulators to target expresion in gene dysregulation analysis.

Usage

DysReg(exp.1, exp.2, tf2tar, 
       de.genes = NULL, de.pval = NULL, de.qval = NULL, de.logFC = NULL, 
       grn.method = 'Boruta', pValue = 0.01, threshold = NULL, 
       ci = 0.95, verbose = TRUE, ...)

Arguments

exp.1	Expression matrix of a special condition. Columns correspond to genes, rows correspond to experiments. The matrix is expected to be already normalized.
exp.2	Expression matrix of an another special condition. Columns correspond to genes, rows correspond to experiments. The matrix is expected to be already normalized.
tf2tar	The prior reference GRN containing TF-target relationships.
de.genes	A dataframe for differential expression genes. If no de.genes offered, DysReg uses the default method limma to implement differential expression analysis. If de.genes offered, The dataframe must include three columns, "GeneSymbol", "high.condition", "de.logFC". "high.condition" means which condition represents high expression level. "de.logFC" is the output logFC from differential expression analysis.
de.pval	The cutoff of pval for filtering differential expression genes. If you don't use this parameter to filter differential expression genes, this parameter could be set as NULL. If you use this parameter to filter differential expression genes, this parameter could be set as a special number, such as 0.05.
de.qval	The cutoff of qval used for filtering differential expression genes. If you don't use this parameter to filter differential expression genes, this parameter could be set as NULL. If you use this parameter to filter differential expression genes, this parameter could be set as a special number, such as 0.05.
de.logFC	The cutoff of absolute logFC used for filtering differential expression genes. If you don't use this parameter to filter differential expression genes, this parameter could be set as NULL. If you use this parameter to filter differential expression genes, this parameter could be set as a special number, such as 0.5. This parameter could be used by combining with de.pval or de.qval.
grn.method	The method used to build conditional GRN, such as 'Boruta', or 'RGBM'.
pValue	Confidence level used in Boruta. Default value should be used.
threshold	The threshould for weight in RGBM.
ci	The confidence invetal of coefficient.
verbose	A logical value indicating whether display the computating progress.
...	Other parameters passed to condiGRN.

Details

DysReg first builds conditional GRNs with feature selection algorithm, where regulatory intensity and its confidence interval of each link is estimated by a de-biased LASSO method. Gene dysregulations were then identified by integrating three properties including differential regulation, differential expression of target, and the consistency between differential regulation and differential expression.

Value

The results of gene dysregulation analysis:

de.genes	The differential expression genes between conditions.
dysreg	The identified gene dysregulations

References

Li Q, Li J, Dai W, et al. Differential regulation analysis reveals dysfunctional regulatory mechanism involving transcription factors and microRNAs in gastric carcinogenesis. Artif Intell Med. 2017, 77, 12-22.

See Also

condiGRN; quantiReg

Examples

data(ExpData)
ExpData[1:5,1:5]

data(tf2tar)
head(tf2tar)

data(ClinData)
head(ClinData)

group.1 <- ClinData$sample[which(ClinData$binaryResponse == 'CR/PR')]
exp.1 <- ExpData[,colnames(ExpData) %in% group.1]

group.2 <- ClinData$sample[which(ClinData$binaryResponse == 'SD/PD')]
exp.2 <- ExpData[,colnames(ExpData) %in% group.2]

dysreg.out <- DysReg(exp.1 = exp.1, exp.2 = exp.2, tf2tar, 
                     de.genes = NULL, de.pval = 0.05, 
                     grn.method = 'Boruta', 
                     pValue = 0.01, ci = 0.90, verbose = T)

dysreg <- dysreg.out$dysreg
head(dysreg)

SCBIT-YYLab/DysRegSig documentation built on July 19, 2021, 4:38 a.m.